Taxonomy of social network data types

Abstract

Online social networks (OSNs) have become an integral part of social interaction and communication between people. Reasons include the ubiquity of OSNs that is offered through mobile devices and the possibility to bridge spatial and temporal communication boundaries. However, several researchers have raised privacy concerns due to the large amount of user data shared on OSNs. Yet, despite the large body of research addressing OSN privacy issues, little differentiation of data types on social network sites is made and a generally accepted classification and terminology for such data is missing. The lack of a terminology impedes comparability of related work and discussions among researchers, especially in the case of privacy implications of different data types. To overcome these shortcomings, this paper develops a well-founded terminology based on a thorough literature analysis and a conceptualization of typical OSN user activities. The terminology is organized hierarchically resulting in a taxonomy of data types. The paper furthermore discusses and develops a metric to assess the privacy relevance of different data types. Finally, the taxonomy is applied to the five major OSNs to evaluate its generalizability.

A Privacy-Enhanced Friending Approach for Users on Multiple Online Social Networks

“…The level of sensitivity of attributes is obtained from the user's perspective and indicates how concerned a user is about his or her information being shared publicly. Regarding a user's attributes, there are various classifications by different authors [24][25][26]. Arnes et al [24] classified users' profile data into three categories consisting of mandatory, extended, and personal data.…”

“…Arnes et al [24] classified users' profile data into three categories consisting of mandatory, extended, and personal data. Ritcthammer et al [26] extended the Arnes classification by covering more attributes on social networking sites to classify the data. Ho et al [25] proposed five different groups to classify users' data.…”